The article presents our research in the field of the automatic diagnosis of large intestine diseases on endoscopic video. It focuses on the methods of selecting regions of interest from endoscopic video frames for further analysis by specialized disease recognition algorithms. Four methods of selecting regions of interest have been discussed: a. trivial, b. with the deletion of characteristic, endoscope specific additions to the video, c. with a cropped square area and d. with a cropped circular area. All four methods have been evaluated in terms of their influence on the efficiency of a set of exemplary disease recognition algorithms [ref], which make use of neural networks and Support Vector Machines. Testing and evaluation have been performed by training the classifiers with an extensive cross-validation and further independent evaluation on a separate test set. The tests have been performed in an advanced testing environment, which is based on an extensive database of various cases, collected by the Medical University of Gdańsk. The article is concluded with the test results and their analysis. The most promising methods and algorithms have been singled out and directions for further research aiming at increasing their effectiveness in recognizing large intestine diseases have been specified.
Authors
Additional information
- Category
- Aktywność konferencyjna
- Type
- publikacja w wydawnictwie zbiorowym recenzowanym (także w materiałach konferencyjnych)
- Language
- angielski
- Publication year
- 2012
